World Journal of Oncology, ISSN 1920-4531 print, 1920-454X online, Open Access
Article copyright, the authors; Journal compilation copyright, World J Oncol and Elmer Press Inc
Journal website https://wjon.elmerpub.com

Original Article

Volume 000, Number 000, July 2026, pages 000-000


Prediction of Residual Disease in Unresectable Head and Neck Squamous Cell Carcinoma After Concurrent Chemoradiotherapy

Figures

↓  Figure 1. Study flow diagram. CCRT: concurrent chemoradiotherapy; CT: computed tomography; MRI: magnetic resonance imaging.
Figure 1.
↓  Figure 2. Area under the receiver operating characteristic curve (AuROC). CI: confidence interval.
Figure 2.
↓  Figure 3. Calibration plot demonstrating the agreement between the predicted and observed probabilities of residual disease (RD). AUC shown in the plot is equivalent to AuROC. O:E ratio: observed-to-expected ratio; CITL: calibration-in-the-large.
Figure 3.
↓  Figure 4. Decision curve analysis demonstrating the net clinical benefit of the prediction model for residual disease.
Figure 4.
↓  Figure 5. The predicted probability of residual disease (RD) plotted against the model’s linear predictor (xb), highlighting the threshold used for clinical decision-making. The linear predictor cutoff (xb cut) of −0.1161 corresponds to a predicted probability of 0.471.
Figure 5.

Tables

↓  Table 1. Baseline Characteristics Associated With Treatment Response (Residual Disease vs. Complete Response)
 
Prognostic factorsResidual disease (n = 180), n (%)Complete response (n = 202), n (%)P valueAuROC
HPV testing was performed mainly in patients with oropharyngeal tumors. Histological grade was missing in one patient with residual disease. AuROC: area under the receiver operating characteristic curve; BMI: body mass index; HPV: human papillomavirus; IQR: interquartile range; NLR: neutrophil-to-lymphocyte ratio; SD: standard deviation.
Male sex75 (41.7)91 (45.0)0.5360.516
Age (years), mean ± SD55.9 ± 9.958.4 ± 11.00.0180.585
Comorbidities
  Diabetes mellitus8 (4.4)11 (5.4)0.8140.505
  Hypertension23 (12.8)36 (17.8)0.2020.525
  Dyslipidemia9 (5.0)19 (9.4)0.1170.522
Smoking history119 (66.1)121 (59.9)0.2430.531
Alcohol consumption111 (61.7)122 (60.4)0.8340.506
Tumor site< 0.0010.683
  Supraglottis12 (6.7)50 (24.8)
  Oropharynx72 (40.0)105 (52.0)
  Hypopharynx70 (38.9)35 (17.3)
  Oral cavity26 (14.4)12 (5.9)
Histologic grade0.6980.522
  Well differentiated55 (30.8)70 (34.7)
  Moderately differentiated91 (50.8)95 (47.0)
  Poorly differentiated33 (18.4)37 (18.3)
Tumor stage (T-stage)< 0.0010.723
  T18 (4.4)17 (8.4)
  T218 (10.0)53 (26.2)
  T335 (19.4)86 (42.6)
  T4119 (66.1)46 (22.8)
Nodal stage (N-stage)< 0.0010.689
  N130 (16.7)91 (45.0)
  N2112 (62.2)106 (52.5)
  N338 (21.1)5 (2.5)
Hemoglobin (g/dL), mean ± SD11.2 ± 1.912.2 ± 1.9< 0.0010.641
Albumin (g/dL), mean ± SD3.7 ± 0.53.9 ± 0.40.0010.617
HPV-positive6 (7.6)53 (41.4)0.0010.669
NLR, median (IQR)4.3 (2.7, 7.1)3.0 (1.9, 4.4)< 0.0010.670
Maximum tumor diameter (cm), median (IQR)6.5 (5.7, 7.3)3.5 (2.5, 4.7)< 0.0010.914
Tumor size anteroposterior (cm), mean ± SD4.4 ± 1.82.8 ± 1.3< 0.0010.761
Tumor size transverse (cm), mean ± SD4.5 ± 1.72.7 ± 1.2< 0.0010.793
Tumor size vertical (cm), mean ± SD5.7 ± 2.03.1 ± 1.5< 0.0010.850
Maximum lymph node diameter (cm), median (IQR)2.0 (1.2, 5.0)1.8 (1.0, 3.0)0.1690.563
BMI (kg/m2), mean ± SD18.9 ± 3.120.4 ± 3.9< 0.0010.615
Time to initiation of radiotherapy (days), median (IQR)62.0 (49.5, 84.5)61.0 (52.0, 75.0)0.2170.538
Overall treatment time (days), mean ± SD54.5 ± 10.852.5 ± 18.90.4700.566

 

↓  Table 2. Model Development and Predictors of Residual Disease
 
PredictorsUnivariable OR (95% CI)P valueMultivariable OR (95% CI)P value
Overall multivariable model AuROC was 0.929. BMI: body mass index; CI: confidence interval; NLR: neutrophil-to-lymphocyte ratio; OR: odds ratio.
Tumor site
  Supraglottis1.00 (Reference)-1.00 (Reference)-
  Oropharynx2.85 (1.42, 5.74)0.0031.39 (0.46, 4.20)0.551
  Hypopharynx8.33 (3.93, 17.63)< 0.0013.56 (1.10, 11.40)0.033
  Oral cavity9.02 (3.56, 22.88)< 0.0013.77 (0.95, 14.90)0.058
Maximum tumor diameter (cm)3.47 (2.72, 4.43)< 0.0013.37 (2.59, 4.39)< 0.001
Hemoglobin (g/dL)0.76 (0.68, 0.85)< 0.0010.94 (0.79, 1.11)0.475
NLR1.15 (1.07, 1.22)< 0.0011.02 (0.94, 1.10)0.610
BMI (kg/m2)0.88 (0.83, 0.94)< 0.0010.97 (0.88, 1.06)0.519
Time to initiation of radiotherapy (days)1.01 (1.00, 1.02)0.0051.01 (1.00, 1.02)0.016

 

↓  Table 3. Model Performance and Internal Validation
 
ParametersApparent performance (95% CI)Bootstrap-corrected performance (95% CI)
AuROC: area under the receiver operating characteristic curve; CI: confidence interval; CITL: calibration-in-the-large; O:E ratio: observed-to-expected outcomes ratio.
C-statistic (AuROC)0.929 (0.903, 0.954)0.929 (0.905, 0.956)
O:E ratio1.0001.002 (0.930, 1.070)
CITL0.000 (−0.304, 0.304)−0.004 (−0.324, 0.323)
Calibration slope1.000 (0.801, 1.199)0.997 (0.776, 1.232)
Bootstrap shrinkage-0.997

 

↓  Table 4. Classification Performance at a Predicted Probability Threshold of 0.471
 
Model probability of residual diseaseResidual disease (n = 180), n (% of total cohort)Complete response (n = 202), n (% of total cohort)PPV (%) (95% CI)NPV (%) (95% CI)
Percentages are cell percentages calculated using the total study population (N = 382) as the denominator. CI: confidence interval; NPV: negative predictive value; PPV: positive predictive value.
≥ 0.471160 (41.9)37 (9.7)81.2 (75.1, 86.4)89.2 (83.8, 93.3)
< 0.47120 (5.2)165 (43.2)